ABSTRACT The alkylating agent temozolomide (TMZ) is widely used for the treatment of human glioblastoma multiforme (GBM). TMZ induces a variety of DNA lesions, including O-6-methylguanine (O6meG), N-7-methylguanine (7meG), and N-3-methyladenine (3meA). Many different DNA repair pathways are involved in the repair and tolerance of alkylation damage, including direct repair by methyl-guanine methyltransferase (MGMT), base excision repair (BER), nucleotide excision repair (NER), and mismatch repair (MMR). Although the functions of MGMT, BER, and MMR in alkylation damage repair and tolerance have been extensively studied, the role of NER remains elusive. We recently examined genome-wide repair of 7meG and 3meA lesions using a novel method named as NMP-seq (N-methylpurine sequencing). Our high-resolution repair data revealed a striking asymmetry between the transcribed strand (TS) and the non-transcribed strand (NTS) in the repair of 3meA in BER-deficient yeast cells, with 3meA being repaired more rapidly on the TS. Accordingly, we also found a strong strand asymmetry in adenine mutations across the genome of BER-deficient yeast. These data lead to our hypothesis that the persisting 3meA lesions in BER-deficient cells stall RNA Polymerase II and activate a subpathway of NER known as transcription-coupled nucleotide excision repair (TC-NER), which specifically removes 3meA lesions from the TS of actively transcribed genes. We propose to elucidate the mechanism by which TC-NER repairs 3meA and prevents mutations on the TS of genes in yeast and human cells (Aim 1). Our findings in 3meA repair also led us to investigate whether TC-NER plays a role in the repair and mutagenesis of O6meG. We reanalyzed mutations occurring in TMZ-treated, MGMT-deficient GBM tumors, in which most somatic mutations are associated with TMZ-induced O6meG lesions. Strikingly, our data indicate that the TMZ signature mutations in these tumors are enriched on the NTS. This strand asymmetry was only observed following TMZ treatment but not in initial, untreated tumors. These findings suggest that O6meG lesions are preferentially removed from the TS in MGMT-deficient cancer cells, likely by the TC-NER pathway. We propose to test whether the observed mutational strand asymmetry can be recapitulated in the genetically tractable yeast model system, and characterize the contributions of key TC-NER factors to O6meG mutational strand asymmetry. Additionally, the role of TC-NER in O6meG repair will be characterized in a MGMT-negative GBM cell line, using the recently published XR-seq (Excision Repair Sequencing) method (Aim 2). Recent studies have uncovered significantly altered mutation landscape in GBM tumors post-treatment with TMZ. Our study will generate the first genome-wide O6meG excision repair map, which should provide mechanistic evidence for therapy-driven evolution of the GBM genome.